SCCM Guidelines on Family-Centered Care for Adult ICUs: 2024

Abstract

Rationale

For staff in adult ICUs, providing family-centered care is an essential skill that affects important outcomes for both patients and families. The COVID-19 pandemic placed unprecedented strain on care of ICU families, and practices for family engagement and support are still adjusting.

Objectives

To review updated evidence for family support in adult ICUs, provide clear recommendations, and spotlight optimal family-centered care practices post-pandemic.

Panel Design

The multiprofessional guideline panel of 28 individuals, including family member partners, applied the processes described in the SCCM Standard Operating Procedures Manual to develop and publish evidence-based recommendations in alignment with the Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) approach. Conflict-of-interest policies were strictly followed in all phases of the guidelines, including panel selection, writing, and voting.

Methods

The guidelines consist of 4 content sections: engagement of families, support of family needs, communication support, and support of ICU clinicians providing family-centered care. We conducted systematic reviews for 15 Population, Intervention, Control, and Outcomes (PICO) questions, organized amongst these content sections, to identify the best available evidence. We summarized and assessed the certainty of evidence using the GRADE approach. We used the GRADE evidence-to-decision framework to formulate recommendations as strong or conditional, or as best-practice statements where appropriate. The recommendations were approved using an online vote requiring >80% agreement of voting panel members to pass.

Results

Our panel issued 17 statements related to optimal family-centered care in adult ICUs, including one strong recommendation, 14 conditional recommendations, and two best-practice statements. We reaffirmed the critical importance of liberalized family presence policies as default practice when possible and suggested options for family attendance on rounds and participation in bedside care. We suggested that ICUs provide support for families in the form of educational programs; ICU diaries; and mental health, bereavement, and spiritual support. We suggested the importance of providing structured communication for families and communication training for clinicians but did not recommend for or against any specific clinician-facing tools for family support or decision aids, based on current available evidence. We recommended that adult ICUs implement practices to systematically identify and reduce barriers to equitable critical care delivery for families and suggested that programs designed to support the wellbeing of clinicians responsible for family support be developed.

Conclusions

Our guideline panel achieved consensus regarding recommendations and best practices for family-centered care in adult ICUs.

INTRODUCTION

Providing care in the intensive care unit (ICU) that is both patient-centered and family-centered is essential for clinical excellence. Care of families is a core component of high-quality care, as family support and engagement influence patients’ outcomes^1,2, and as ICU experiences have lasting impacts on family members themselves. ³

The Society of Critical Care Medicine (SCCM) has released two prior sets of guidelines for family support^4,5, the most recent in 2017. Following these guidelines, unprecedented momentum began for ICU family support and engagement initiatives⁶, only for the world to endure the global coronavirus disease (COVID-19) pandemic and experience the profoundly negative effects it had on ICU family presence⁷ and staff burnout. ⁸ In this post-pandemic era, this third iteration of the guidelines seeks to review an updated evidence base focused on family engagement in adult ICUs, provide clear recommendations, and spotlight optimal family-centered care practices at a time when they have never been more important.

METHODOLOGY

Question selection and outcome prioritization

Each guideline topic was structured in PICO (population, intervention, comparison, outcome) format. A proposed list of PICO questions was developed by the writing panel chairs and co-chairs (Supplemental Digital Content, SDC 1) and refined with the assistance of the methodologists. These were circulated to the entire panel, who had an opportunity to suggest modifications. A final list of 15 PICO questions (i.e., the maximum number for a single SCCM guideline at the beginning of guideline development) was approved by the chairs (SDC 2). Following GRADE guidance, the panel then developed a list of outcomes, with each outcome rated as “critical,”(mean rating 7–9) “important,” (mean 4–6) or of “limited importance”(mean ≤ 3) to decision-making; considering the perspectives of patients, families, and ICU staff (SDC 3). ⁹

Search strategy and study selection

With input from the methodology team, a medical librarian conducted peer-reviewed systematic searches of article databases and clinical trial registries from database inception until 17 June 2021. These were subsequently updated to March 2023, including articles identified as protocols or registered and subsequently published, as well as articles identified by the expert panel (SDC 4 and SDC 5). Results were uploaded to COVIDENCE for screening. Each abstract was screened in duplicate by two screeners, and any reference marked by either screener as potentially relevant was advanced for full-text screening. Full text screening was done in duplicate, with the two methodologists (SJWO or KL) making a final decision of study eligibility in the event of disagreement between the two screeners. Studies of multifaceted interventions that applied to more than one PICO were permitted as evidence for each relevant PICO.

Data abstraction, analysis, and evidence summaries

Using a piloted data abstraction spreadsheet, the methodologists conducted data abstraction and assessed risk of bias using the Cochrane Risk of Bias tool for randomized controlled trials (RCTs). ¹⁰ Meta-analyses of evidence for each PICO were conducted using RevMan 5.4¹¹, with inclusion of non-RCT data into the meta-analyses allowed on an individual question-by-question basis, if insufficient randomized evidence was available for critical outcomes. For all outcomes, the methodologists compared fixed- and random-effects estimates, preferentially using random-effects unless there were concerns that small-study effects may have been impacting the pooled estimate. ¹² For dichotomous variables, we reported relative risk (RR) and absolute risk difference (RD). For continuous variables, we reported mean difference (MD), or standardized mean difference (SMD), as appropriate, each with a corresponding 95% confidence interval (95% CI).

The methodologists developed an evidence summary for each PICO. Qualitative evidence and data otherwise not amenable to meta-analyses were summarized in tables. For each outcome with available data, we rated the certainty of evidence as “high,” “moderate,” “low,” or “very low” in accordance with GRADE practice. ¹³

Development and voting upon recommendations

Members of the panel were assigned to PICO working groups, which reviewed and provided feedback on respective evidence summaries and worked through Evidence-to-Decision (EtD) frameworks to develop draft recommendations. If only indirect evidence was available, the panel had the option of providing general guidance using ungraded best practice statements, if these met GRADE criteria for their use. ¹⁴ The EtDs considered the balance of desirable and undesirable effects on outcomes of importance, the certainty of evidence, resource implications, health equity (disproportionate effects in some groups), feasibility, and acceptability of each intervention. The entire panel then used PanelVoice software to vote upon each recommendation, its justification, implementation issues, and research considerations. Following SCCM practice, >80% agreement of 70% of eligible panel voters was considered consensus for a recommendation. Of note, the use of the terms “clinician” and “staff” in this manuscript reflect that of each individual PICO working group during their respective discussions, but are meant to be broadly inclusive of all ICU care team members.

Implementation tools development

A task force comprised of volunteer members from this writing panel and from the SCCM THRIVE Patient and Family Support Committee adapted a gap analysis toolkit and listing of widely available family-centered care resources from the 2017 guidelines¹⁵ to create a new set of tools, once again designed to assist frontline leaders of ICUs with implementing our updated recommendations in their local environments. This updated toolkit is provided on the SCCM guidelines web page (www.sccm.org/guidelines).

RECOMMENDATIONS

All recommendations and best practice statements are summarized in Table 1.

Table 1.

Summary of recommendations.

Recommendation	Strength of recommendation	Certainty of evidence
General engagement of families
We recommend liberalized ICU family presence policies as the default practice in intensive care units.	Strong	Low
We suggest offering the option of families being present on rounds in the ICU.	Conditional	Very low
We suggest offering family participation in bedside care.	Conditional	Very low
We suggest offering family presence during resuscitation, with an assigned staff member to support the family. There are not sufficient data for a statement regarding family presence during invasive bedside procedures.	Conditional	Low
General support of family needs
We suggest providing educational programs for families of ICU patients to orient them to the ICU environment, ICU team, and ICU concepts.	Conditional	Moderate
We make no suggestion for or against specific tools relevant to family-centered ICU care designed for clinical teams.	Conditional	Low
We suggest providing ICU diaries for families of ICU patients.	Conditional	Low
We make no suggestion for or against any specific decision-making support tools for families of ICU patients.	Conditional	Low
We suggest providing bereavement support to families of patients who have died in the ICU. There are not sufficient data to formally recommend a specific intervention. Interventions described in the literature include brochures or booklets; condolence letters; and meetings with specific care teams, such as palliative care, psychologists, or specially trained nurses.	Conditional	Low
We suggest identifying and supporting the mental health and psychological needs of families of ICU patients. There are not sufficient data to formally recommend a specific intervention. Interventions described in the literature include meetings with clinicians, psychoeducation, and coping skills training.	Conditional	Low
We suggest identifying and supporting the spiritual needs of families of ICU patients. There are not sufficient data to formally recommend a specific intervention. Interventions described in the literature include proactively making families aware of the option for chaplain visits.	Conditional	Low
We recommend ICUs use family support zones or incorporate supportive features to meet family needs during patients’ ICU stays. Family needs described in the literature include rest spaces, areas for personal care, spaces to interact with staff for sensitive and confidential discussions, and room to sit comfortably at the bedside of the patient.	Best practice statement	Best practice statement
We recommend that ICUs implement practices to systematically identify and reduce barriers so as to promote equitable critical care delivery for patients’ families. Barriers described in the literature include language, varied cultural understandings and beliefs, and varied expectations around health and critical care delivery.	Best practice statement	Best practice statement
Communication support for families
We suggest using standardized approaches for interdisciplinary family conferences and facilitation of communication in ICUs. Interventions include the use of specialized ICU staff who facilitate communication with families and regularly scheduled structured family meetings, including clinicians’ use of the VALUE mnemonic (Value family statements, Acknowledge emotions, Listen, Understand the patient as a person, Elicit Questions).	Conditional	Low
We suggest ICUs provide communication skills training to clinicians, if local resources permit.	Conditional	Low
We suggest critical care trainees participate in high-fidelity (e.g., standardized actor) simulation communication education training programs. There are not sufficient data to formally recommend a specific training program.	Conditional	Low
Support for ICU clinicians responsible for family-centered care
We suggest structured programs to support clinicians in promoting the delivery of family-centered ICU care. Programs described in the literature have typically been multifaceted quality-improvement programs focused on family outcomes. There are not sufficient data to formally recommend a specific program, particularly with regards to clinician burnout, stress, and psychological outcomes.	Conditional	Low

General engagement of families

General family presence within ICUs

Recommendation.

We recommend liberalized ICU family presence policies as the default practice in intensive care units (strong recommendation, low certainty evidence).

Rationale.

In addition to 17 observational and quasi-experimental studies, our panel identified 8 RCTs^16–23, which varied greatly in terms of how “liberalized” vs. “restrictive” family presence were defined and operationalized. The largest RCT was the Brazilian ICU Visits Study¹⁹, a cluster-RCT of over 1000 family members which found no difference between liberalized vs. restrictive groups with respect to a primary outcome of patient delirium, but which found better family anxiety and depression and family satisfaction in the liberalized group as secondary outcomes. In our meta-analyses, we found that liberalized policies demonstrated increased family satisfaction, possibly increased patient satisfaction, and reduced family and patient symptoms of anxiety and depression, measured using a variety of validated instruments (SDC 6). We did not identify any undesirable effects of family presence, with no impact of liberalized presence upon patient mortality, length of stay, and ICU-acquired infections—although all included studies were conducted prior to the COVID-19 pandemic. Similarly, we did not identify any negative effects upon clinician satisfaction, anxiety, burnout, or conflict with family; however, this evidence was of lower certainty due to short length of follow-up.

The COVID-19 pandemic brought profound challenges; and we acknowledge that policies to restrict ICU family presence may be justified if presence has a disproportionately negative impact upon patient care, staff workload, or burnout. ²⁴ However, our panel viewed this recommendation as one of particular importance, given the core impact of family presence on other aspects of family engagement; the humanistic importance of decreasing family anxiety and depression and improving family satisfaction; and the increased availability of RCT data compared to the last iteration of these guidelines⁵. Despite the need for future study, particularly with regards to pandemic-era outcomes²⁵, our panel voted in favor of a recommendation that specifically stated liberalized family presence as the default option in ICUs, as opposed to the default being more restrictive. While the degree of “liberalized” family presence in policies may be in question, defaulting to more flexible family presence hours, whenever possible from a safety standpoint, promotes equity in allowing families more options for being in the ICU and additionally possibly promotes clinician-family trust in decision-making. ²⁶

Recommendation.

We suggest offering the option of families being present on rounds in the ICU (conditional recommendation, very low certainty evidence).

Rationale.

Only one small RCT evaluated families being present on rounds in a cardiac ICU. This evidence was rated down for indirectness, as the patients in the study were non-ventilated and well enough to participate in rounds themselves, unlike most ICU patients. ²⁷ All other studies reviewed were observational studies, including some qualitative studies assessing family and staff perceptions of family participation. Very low certainty evidence suggested family participation in rounds might increase family self-reported knowledge of patients’ care plans and improve ratings of quality of communication, as well as improve some patients’ satisfaction with care and symptoms of anxiety (SDC 7). Amongst those studies that tracked clinician outcomes, no adverse effects of the intervention on staff workload, time spent on rounds per patient, or staff satisfaction were found. Of note, qualitative data did reveal implementation challenges with family presence on rounds; including coordinating logistics, potentially revealing disagreements amongst care team members, and possibly overloading family members with information. ²⁸

The panel made a conditional recommendation on this limited evidence, judging that it demonstrates that (1) family presence on rounds is feasible; and (2) carefully implemented, it may have some desirable effects and has no clear evidence of undesirable effects. The panel noted that the magnitudes of both positive and negative effects are still currently uncertain. Furthermore, the panel had uncertainty regarding the impact that family presence on rounds would have upon equity of care. Some families may benefit greatly from active participation with clinicians on rounds, while others may not have the resources (or time) to be able to attend rounds consistently and may thus be at risk for inadvertently receiving less communication. ²⁹ Ultimately, how best to protocolize incorporation of families into ICU rounding structures is an important question that would benefit from additional trials.

Participation in bedside care

Recommendation.

We suggest offering family participation in bedside care (conditional recommendation, very low certainty evidence).

Rationale.

The evidence base for this recommendation included 5 small RCTs^30–34 and 3 observational studies; heterogeneity in both design and outcome measures made interpretation challenging. Family participation in bedside care may improve family symptoms of anxiety. While observational evidence suggests it may improve family satisfaction with care and family symptoms of post-traumatic stress disorder (PTSD), these effects were not seen in the RCTs (SDC 8). Studies did not find clear benefit in family depression symptoms or reduced patient delirium. However, studies also did not identify any harmful aspects to patients of offering families the option to participate.

Overall, the panel judged that offering family participation in bedside care may benefit some families with respect to mental health symptoms and likely has few undesirable effects, so long as the participation is within the family’s interest and ability. Thus, the panel made a conditional recommendation for offering families the option to participate in bedside care. The primary barrier to participation is feasibility, as it requires skilled ICU staff with sufficient resources to train and oversee families with regards to their activities and institutional support. Staff may be uncomfortable with family participation in care, depending on their previous experiences and the details of implementation. ³⁵ Family willingness and ability to participate in care may vary over time. Training programs that better elucidate assessment of patient and family preferences and optimal implementation format(s), with careful assessment of outcomes, are needed.

Family presence during cardiopulmonary resuscitation or bedside procedures

Recommendation.

We suggest offering family presence during resuscitation, with an assigned staff member to support the family (conditional recommendation, low certainty evidence). There are not sufficient data for a statement regarding family presence during invasive bedside procedures.

Rationale.

The evidence base for this recommendation consisted of 3 RCTs^36–38 and 5 observational studies. Our meta-analyses revealed no impact on patient survival but limited reductions in family members’ symptoms of anxiety, depression, and complicated grief (SDC 9). However, benefits in family satisfaction, family PTSD, or staff outcomes were not detected. Our panel thus based this recommendation on a consensus that offering family presence during resuscitation may benefit some family members and likely has minimal effect on mortality or resuscitation efforts. The panel noted that the RCTs included the use of trained support staff for families, which may have minimized any negative impacts of family presence. These interventions may increase staff self-efficacy, if institutional resources allow for proper training and presence of additional staff, including chaplaincy. A family member’s choice to participate is important, as not all situations are the same; it is important for clinicians to offer families the option to be excused or leave the room. From an equity standpoint, there are limited data on the impact of family presence for historically underserved populations. Research moving forward should assess the benefits of witnessing a resuscitation, possible unknown harms, and the optimal structure for training support staff.

The panel judged there was too little evidence to make a recommendation on family presence during other invasive bedside procedures. One quasi-experimental study found that families wanted to be present during procedures but were rarely asked, and family presence had no adverse effect on PTSD symptoms. ³⁹ Another study suggested that trainees’ educational experiences were not negatively impacted and stress did not increase when families were present for procedures. ⁴⁰

General support of family needs

Educational programs for families

Recommendation.

We suggest providing educational programs for families of ICU patients to orient them to the ICU environment, ICU team, and ICU concepts (conditional recommendation, moderate certainty evidence).

Rationale.

Our panel reviewed 10 RCTs^41–50 from the US, France, and China; testing a variety of programs: informational leaflets, electronic tablets with online orientation materials, a guide advising families on questions to ask the clinical team, one-on-one orientation sessions with an ICU staff member, and an ICU tour. Taken together, these studies did demonstrate with moderate certainty a variety of desirable effects, including improvements in family satisfaction (overall), family satisfaction specifically with care received, family anxiety symptoms, patient satisfaction with care, patient mental health, and patient quality of life (SDC 10). A combined analysis of those trials that studied impact on family comprehension was not able to show significant improvement on this outcome, although this finding had very low certainty due to variability in individual study results. Also, our meta-analyses did not reveal an impact of these interventions on patient length of stay. Overall, our panel concluded the undesirable effects of these interventions to be trivial.

Although the balance of effects favored providing educational programs, we stopped short of issuing a strong recommendation in favor of these interventions; given the lack of studies examining the cost-effectiveness, feasibility, and impact on equity of the varied strategies that have thus far been tested. Of note, one study that did include nursing satisfaction as an outcome was able to show a significant benefit of a one-on-one session with a clinician. Future studies to further examine how best to balance costs, overcome language and health literacy barriers, promote clinician outcomes, and tailor interventions to family populations with specific needs would be helpful for maximizing implementation.

Clinician-facing tools that assist with provision of family-centered care

Recommendation.

We make no suggestion for or against specific tools relevant to family-centered ICU care designed for clinical teams (conditional recommendation, low certainty evidence).

Remarks.

Tools “designed for clinical teams” (as opposed to “designed for families”) refers to interventions intended to be used directly by ICU professionals to enhance their ability and consistency in providing family-centered care. Examples include checklists and protocols designed to direct the clinical team’s attention towards family-centered communication^51,52, shared decision-making^53,54, or patient and family palliative care needs⁵⁵.

Rationale.

The majority of the 8 studies^51–58 we reviewed for this recommendation utilized non-randomized quasi-experimental methodologies. Qualitative data suggested that participating clinical teams were supportive of such interventions. However, our meta-analyses did not demonstrate intervention effects on family satisfaction, family psychological symptoms, patient length of stay, or patient mortality (SDC 11). Of note, one single-center study⁵⁶ evaluated a multi-faceted intervention with both clinician-facing (e.g., safety checklist, blog for team members, care plan worksheet) and patient-facing components (e.g., electronic messaging of care team). This single study did demonstrate improvement in overall hospital ratings amongst a small subset of participants who completed the Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS). However, our panel felt that generalizing patient HCAHPS scores to the ICU population is inadvisable due to the survey’s inherent selection bias and content focus on the overall hospital experience.

Given the general benefit of a structured and systematic approach to patient care, our panel debated suggesting a conditional recommendation in favor of clinician-facing tools, but ultimately decided against such a statement. We acknowledge that the lack of evidence that these tools have a positive impact on families is quite possibly due to the imprecision of family outcome measurement instruments. Current instruments for measuring family satisfaction can suffer from poor discrimination and responsivity. Ultimately, while interventions designed for clinicians appeared to be feasible to implement in some ICUs without undesirable effects, our panel was uncertain whether any specific interventions represented optimal utilization of limited resources, especially when compared to other family-centered care strategies that are directly family-facing.

ICU diaries

Recommendation.

We suggest providing ICU diaries for families of ICU patients (conditional recommendation, low certainty evidence).

Rationale.

Protocols of studies of ICU diaries have used a variety of multidisciplinary complex behavioral approaches, including having family members write diary entries at a variety of time intervals, having diaries written by both ICU staff and family, having an ICU diary team write and read the diary with patients, and having the diary filled out by the care team in conjunction with a psychoeducational program. Amongst the studies reviewed for this recommendation, 12 were RCTs^59–70, with multifaceted interventions and outcomes. Although the most prominent recent French RCT of ICU diaries kept by clinicians and family members for mechanically ventilated patients did not demonstrate an impact on patient PTSD symptoms⁵⁹, our overall meta-analyses found significant beneficial associations between ICU diary protocols and patient PTSD, albeit with low certainty (SDC 12). We found no differences for any other patient outcomes, including delirium, anxiety, depression, quality of life, mortality, and length of stay. For family outcomes, we found no impact of the intervention on anxiety, depression, or PTSD.

Our panel judged the evidence likely in favor of the intervention, due to a reduction of PTSD among patients in the meta-analyses and no clear undesirable effects, although we recognize there is some cost to provide diaries and staff time needed train families and/or staff on protocols. Similar to educational materials provided to families, language barriers may inadvertently create inequities.

Family-facing decision support tools

Recommendation.

We make no suggestion for or against any specific decision-making support tools for families of ICU patients (conditional recommendation, low certainty evidence).

Remarks.

The term “decision support tools” refers to paper-based or multimedia decision aids that supplement family meetings, typically regarding goals-of-care discussions. These aids provide families with information regarding patient prognosis and treatment options, as well as elicit family values. The 2017 guidelines suggested that decision support tools be implemented when possible⁵, primarily based on a single quasi-experimental study of a decision aid for surrogates of patients with prolonged mechanical ventilation⁷¹.

Rationale.

Our recommendation is informed by 7 RCTs^43,72–77 and several non-RCTs, nearly all of which examined aids focused on patient prognosis and goals-of-care decision making. While a French RCT enrolling 90 family members making end-of-life decisions demonstrated lower proportions of PTSD and depression at 3 months amongst those participants randomized to an informational pamphlet about decision-making⁴³, there was no overall effect of decision support tools on standardized scores of anxiety, depression, and PTSD in our RCT meta-analysis. An RCT randomizing 416 surrogates of prolonged mechanical ventilation patients to a web-based decision aid about treatment options did demonstrate a reduction in decisional conflict in its intervention arm as a secondary outcome⁷², but no overall effect of intervention on decisional conflict was seen in our meta-analysis. Our meta-analyses did not reveal any differences in effects regarding other family (anxiety, comprehension), patient (ICU length of stay, mortality, discharge disposition), clinician (quality of communication, nursing distress), and relational (family-clinician prognostic concordance) outcomes that current studies variably included (SDC 13). We thus judged the current evidence for both desirable and undesirable effects to be of trivial magnitude, with low certainty.

In forming this recommendation, we noted that relevant studies found that family members and clinicians viewed decision aids as acceptable and feasible for use. However, selecting study outcomes that demonstrate a complete picture of a decision aid’s effects is challenging, since there is uncertainty about the importance of existing available outcomes measures and of potentially unmeasured outcomes.

Bereavement support

Recommendation.

We suggest providing bereavement support to families of patients who have died in the ICU (conditional recommendation, low certainty evidence). There are not sufficient data to formally recommend a specific intervention. Interventions described in the literature include brochures or booklets; condolence letters; and meetings with specific care teams, such as palliative care, psychologists, or specially trained nurses.

Rationale.

Our recommendation is informed by 8 RCTs^78–85, with low quality evidence, little evidence for potential harm, and perceived high potential for benefit. Outcomes varied highly between trials and include family anxiety, depression, post-traumatic stress disorder, grief, comprehension of care, satisfaction with care, and quality of communication. Different meta-analyses assessing family satisfaction with care, family satisfaction with decision making, and general family satisfaction showed no statistically significant improvements associated with bereavement support initiatives (SDC 14). There was a non-significant trend between control interventions and high general family satisfaction, though this could be due to other innate or related factors characteristic to these types of studies. ^80,86,87 The results of meta-analyses between bereavement support and family member anxiety varied between no statistically significant association (proportion of families with anxiety at first follow-up) to a trend supporting the intervention (family anxiety at one month) to statistically significant benefit associated with the intervention (family anxiety scores at six months and proportion of families with anxiety at six months).

Since inception, critical care units have been a site of death for many patients and their family members and consequently, both palliative and end-of-life care are in the core skillset for interprofessional and multidisciplinary critical care practitioners. Provision of evidence-informed and effective bereavement support for the families of ICU decedents is critical for not only those family members but also critical care practitioners themselves. ^88–90 Although evidence is limited in both quantity and quality, we found the lack of significant harm and the potential for associated benefits of bereavement support to families to be worthy of recommendation. Bereavement support should be tailored to local considerations and dependent on baseline factors such as resources, unit-based champions for bereavement support, and general staff burden.

Family mental health and psychological needs

Recommendation.

We suggest identifying and supporting the mental health and psychological needs of families of ICU patients (conditional recommendation, low certainty evidence). There are not sufficient data to formally recommend a specific intervention. Interventions described in the literature include meetings with clinicians, psychoeducation, and coping skills training.

Rationale.

The 14 RCTs^{49,64,79,81,82,91–99} reviewed for this recommendation overlapped with those reviewed for other PICO questions and examined a wide variety of multidisciplinary, complex behavioral interventions that targeted mental health outcomes amongst families, including: nurse-led assessments of mental health; awareness-focused training; coping skills training; internet-based cognitive-behavioural writing therapy; “facilitated sensemaking”, or assisting families multidimensionally with adapting to the ICU environment and making meaning of their situations; post-ICU self-help rehabilitation; and psychoeducation. Our meta-analyses found significant associations between these interventions and the proportion of patients with PTSD symptoms, as well as improved family satisfaction and nursing satisfaction with low certainty (SDC 15). Aside from conflicting evidence on family anxiety, there were no collective differences found for any other outcomes; including depression and PTSD symptoms; family quality-of-life; patient ICU length of stay; patient mortality; patient anxiety and depression symptoms; and patient quality of life. No increase in undesirable effects was found for any outcome; however, it is possible that not all undesirable effects were measured.

Overall, our panel felt that the available evidence favored interventions supporting ICU family mental health, due to reduction of patient PTSD and improved family satisfaction. With regards to implementation, hidden costs of such programs to target family mental health may exist; including nurse training, cost of psychologists to administer psychological interventions, and costs of screening patients and their families for identifying high-yield participants. Work to promote scalability of such interventions to a variety of ICU environments will be important for ensuring equity of support.

Family spiritual needs

Recommendation.

We suggest identifying and supporting the spiritual needs of families of ICU patients (conditional recommendation, low certainty evidence). There are not sufficient data to formally recommend a specific intervention. Interventions described in the literature include proactively making families aware of the option for chaplain visits.

Rationale.

While RCTs are limited, interventions evaluated^100–104 in the literature to better identify and address the spiritual needs of families of critically ill patients include but are not limited to: incorporation of a chaplain or a “chaplain navigator”; multi-disciplinary, complex behavioral interventions incorporating elements such as spirituality and/or palliative care education, local champions, local detailing, standardized order sets, and/or audit with feedback; and chaplain-developed spiritual communication cards disseminated to critically ill patients and their families. One recent RCT enrolling 192 critically ill patient and surrogate dyads testing a protocolized intervention of proactive chaplain support was able to demonstrate benefits with surrogate mental health symptoms and reduced decisional conflict. ¹⁰² Existing data do not suggest any detrimental effects. Upon our meta-analyses of two available RCTs^100,102 and multiple quasi-experimental studies of ICU spiritual support, we found no overall associations between increased spiritual support and length of ICU stay or in-hospital mortality (SDC 16).

While evidence supports that a high proportion (>70%) of surrogate decisions makers endorse religion or spirituality as fairly or very important in their life, the topic is rarely introduced by critical care practitioners or incorporated into family meetings for critically ill patients. ^105,106 While not all families will value spiritual support, a high proportion of families strongly value it. Supplementing the available RCT data, our review of non-RCT data supported the ideas that providing spiritual support can improve how people feel, including feeling better about dealing with the hospitalization, feeling more at peace, and feeling more connected with what is sacred. Costs and sustainability associated with provision of supplemental spiritual care are unclear, though existing data support that provision is both feasible in many critical care environments and often dependent on hospital- or critical care unit-related chaplaincy services.

Physical ICU design

Best practice statement.

We recommend ICUs use family support zones or incorporate supportive features to meet family needs during patients’ ICU stays. Family needs described in the literature include rest spaces, areas for personal care, spaces to interact with staff for sensitive and confidential discussions, and room to sit comfortably at the bedside of the patient.

Rationale.

We identified 20 observational studies of family experiences with ICU design (SDC 17). The literature assessing the impact of single-bed rooms largely consists of quasi-experimental studies in which ICUs were relocated from locations with multi-bed design to ones with single beds. Multiple studies report pre-post improvements in family satisfaction measures due to single bed rooms. ^107–109 Provision of family support zones with sleeping surfaces increase the proportion of families staying at the bedside overnight, an observation suggesting these zones are valued by some families. ¹¹⁰ Among other room features in our literature search reported by families to improve satisfaction or increase presence or engagement were larger room size, seating areas, natural light, and low noise.

Several narrative studies summarized in SDC 17 assessed common areas for families of ICU patients. Qualities that better supported families included adequate seating, availability of vending machines, close proximity to the ICU, natural light, and presence of natural plants. Families also voiced the need for consultation rooms for meetings in which privacy was desired or for which family preference was not to talk within patients’ rooms. ¹¹¹ Given this indirect evidence, the panel made a best practice statement per GRADE guidance supporting the inclusion of family areas in room design and supportive features for families in ICUs.

Equity and inclusion

Best practice statement.

We recommend that ICUs implement practices to systematically identify and reduce barriers so as to promote equitable critical care delivery for patients’ families. Barriers described in the literature include language, varied cultural understandings and beliefs, and varied expectations around health and critical care delivery.

Rationale.

We found 15 observational studies of varied design spanning over two decades that details multiple and multi-faceted disparities in critical care delivery and outcomes across varying populations and settings (SDC 18). These include findings such as: family members at lower income hospitals reporting a higher prevalence of anxiety and depression with lower quality of life¹¹²; self-identified family members of African American patients reporting lower received scores of Professional Support as compared to an ideal score¹¹³; and an ethnographic study in a multi-ethnic urban hospital noting several challenges related to delivery of bad news, identification of legal representatives, perceptions related to comprehension, disclosure of negative information, language barriers, and distrust of clinicians by families. ¹¹⁴ There were insufficient data to support the use of any one method to reduce care inequities.

Identifying and mitigating barriers to equity and inclusion is inherently a positive activity, is unlikely to have substantial costs, and will likely enhance communication and understanding between patients, family members, and the ICU care team. Our panel voted on this best practice statement to support the systematic identification and reduction of barriers to equitable care for families in the ICU.

Communication support for families

ICU family conferences and communication facilitation

Recommendation.

We suggest using standardized approaches for interdisciplinary family conferences and facilitation of communication in ICUs (conditional recommendation, low certainty evidence). Interventions include the use of specialized ICU staff who facilitate communication with families and regularly scheduled structured family meetings, including clinicians’ use of the VALUE mnemonic (Value family statements, Acknowledge emotions, Listen, Understand the patient as a person, Elicit Questions).

Rationale.

With regards to ICU protocols involving specialized communication facilitators, while multiple RCTs and quasi-experimental studies were included in our literature review, our recommendation was informed by two multi-center RCTs in particular. ^115,116 In these trials, reduction in ICU length of stay was found among decedents, and quality of communication was higher among families working with facilitators. ¹¹⁶ However, no differences were detected for family anxiety or PTSD symptoms in either. For these trials, it is unclear which elements of their multifaceted intervention protocols affected the observed differences in outcome. We noted in a separate single-center RCT that length of stay was longer in the facilitator group. ¹⁰⁰ Future studies will need to better delineate training and implementation of the ICU communication facilitator role.

Variability among many studies examining other standardized communication protocols limited meaningful meta-analyses (SDC 19). However, taken together, the literature supports the principle that families benefit from timely and structured communication approaches, with the first family meeting generally recommended within 72 hours of admission. ¹¹⁷ We noted that the 2017 guidelines⁵ supported use of the VALUE mnemonic for patients with poor prognosis and routine interdisciplinary conferences based on RCT evidence, and we continue recommending use of this structure for family meetings after updated literature review. ^83,118,119 Regarding implementation of structured communication methods, we observed benefit from clinician training. Methods of this included communication training with internet-based tools⁵⁶ and structured family meetings with a “train the trainer” component. ¹²⁰

Communication techniques and training

Recommendation.

We suggest ICUs provide communication skills training to clinicians, if local resources permit (conditional recommendation, low certainty evidence).

Recommendation.

We suggest critical care trainees participate in high-fidelity (e.g., standardized actor) simulation communication education training programs (conditional recommendation, low certainty evidence). There are not sufficient data to formally recommend a specific training program.

Rationale.

While the evidence base for these two recommendations included 17 RCTs^{47,83,116,121–134} and numerous observational studies of varied design, heterogeneity in study design, population, and in training approaches (low vs. high fidelity) made recommendations regarding specific training programs challenging. Training populations included mixed ICU staff, nurses, nursing trainees, and physician trainees at both the resident and fellow levels. Evidence for clinical effects of education were derived primarily from low-fidelity (i.e., non-simulated) education of ICU staff (SDC 20). ICU staff education programs likely result in improved quality of communication and may result in improvements in process outcomes (e.g., time to family meetings, documentation of code status) and clinical outcomes (e.g., reductions in length of stay). Staff education may also result in reductions in family member symptoms of anxiety, depression, and PTSD. Studies of communication training programs for learners, focused on high-fidelity (simulation-based) educational programs, mostly evaluated self-efficacy and adherence to communication checklists, and appeared to be effective without undesirable effects. A large multi-center RCT which enrolled 472 internal medicine trainees and nurse practitioners was the exception and evaluated clinical outcomes, suggesting little effect, other than a small increase in depressive symptoms of patients after the training program. ¹²²

Given the heterogeneity between clinical team and learner educational programs, the panel made two separate recommendations from the original single PICO (SDC 2). While the certainty of evidence for most outcomes is low, these training programs appear to improve quality of communication and some process outcomes. The primary constraint is the utilization of resources associated with both, as some of the training programs studied include the use of proprietary materials, and likely require efforts to set up and sustain. Opportunity exists for the development of future communication training programs to better care for historically underserved populations, using standardized approaches to offset potential latent biases.

Support for ICU clinicians responsible for family-centered care

Recommendation.

We suggest structured programs to support clinicians in promoting the delivery of family-centered ICU care (conditional recommendation, low certainty evidence). Programs described in the literature have typically been multifaceted quality-improvement programs focused on family outcomes. There are not sufficient data to formally recommend a specific program, particularly with regards to clinician burnout, stress, and psychological outcomes.

Rationale.

We found a wide spectrum of programmatic approaches to promoting family-centered care, by observational, quasi-experimental, and RCT methodologies—but very few that included important clinician outcomes, such as burnout, stress, and mental health. Studies to date have been much less focused specifically on supporting clinicians and improving these outcomes. ^135,136 The types of programs that the literature search for this recommendation’s PICO (SDC 2) returned were mostly multifaceted quality-improvement programs, including many overlapping with literature reviewed for prior PICOs: communication training, implementing checklists, meeting algorithms, and electronic medical record and mobile applications (SDC 21). With regards to outcomes that were examined, these programs did improve family-perceived patient-centeredness, quality of communication, and satisfaction. However, particularly in the wake of the COVID-19 pandemic, our panel believed there to be a critical need for developing programs that support ICU clinicians directly in their efforts to provide compassionate and empathetic care to patients’ families, focused on supporting clinicians’ resilience and mental health. ^137,138

RESEARCH AGENDA.

In addition to the opportunities for further research that our panel has highlighted within our rationale for each individual statement above, we acknowledge that widespread restrictive family presence policies of the COVID-19 pandemic era resulted in phone communication and especially video conferencing with ICU families becoming more commonplace. ⁷ Many of our recommendations in these guidelines are ideally implemented with families at patients’ bedsides, and having a liberalized family presence as a default policy is the only recommendation our panel rated as “strong”. However, research to understand effective protocols for supporting and engaging families remotely and identifying interventions that positively affect family (including children), patient, and clinician outcomes—even when families themselves are not present at the bedside—will be important for preparing for the next event that necessitates such measures.